AN ACOUSTIC EVALUATION OF ELECTRONIC
Pestology Centre, Department of Biological Sciences, Simon Fraser University, Burnaby, B.C.
V5A 1S6, Canada
ABSTRACT. An analysis of the acoustic
output of 4, commercial devices claimed to
repel mosquitoes showed fundamental fre-
quencies between 1.95 and 5 kHz and many
harmonics extending into the ultrasonic range.
The development and marketing of
several small sound generators intended
to repel female mosquitoes followed the
publication in an electronic hobbyist's
journal of an article entitled "Build the
Bug-Shoo" (Greenlee 1970). This
pocket-sized device was claimed to be ef-
fective because "The female mosquito is
thought to be repelled by the same sound
that attracts the male," The sound that
attracts the male, in species that mate in
swarms, is generated by the wing move-
ment of the female (Roth 1948, Belton
and Costello 1979) and although this
sound is complex, only the fundamental
frequency of the wingbeat — a sine
wave — is essential to attract males
(Wishart and Riordan 1959).
Another statement in the article is that
"the male mosquito is attracted by a
humming noise at a frequency of 2000
Hz" (cycles per second). This is
incorrect — the males of most pest species
in North America are attracted by fre-
quencies of 200-350 Hz, the wingbeat
frequencies of the corresponding females
(Belton 1967, Belton and Costello 1979).
Set at the frequency he suggests,
Greenlee's device produces a sound 6-10
times too high to duplicate the buzz of a
The Bug-Shoo, a unijunction-transistor
oscillator driving a miniature earphone,
was built in 1972. Its output waveform
consisted of a train of pulses that could be
varied over a wide range of frequencies.
Its sound is, therefore, complex and
None of them had significant repellent effects
on laboratory bred Aedes aegypti or Cukx pipiens
or on field populat ions, mostly of At. hexodon-
contains harmonics whose frequencies
depend on the pulse rate. Because it
could be tuned over a wide range of fre-
quencies a comprehensive analysis of this
device was not practical.
The 4 commercial sound generators
tested all had fixed frequency outputs.
Advertising for one of them claimed that
its output contained ultrasonic frequen-
cies that simulated an echolocating bat
and that this would repel biting insects
(Costello and Brust 1976). Another
generator Was claimed by its manufactur-
ers to simulate the buzz of a mate mos-
quito which they suggested would repel
mated females. Three of the generators
appear to be based on the unijunction
oscillator of Greenlee and therefore
should have pulsed outputs but the only
descriptions in the literature of the
acoustic output of commercial mosquito
"repellers" stated that they were sine
waves (Kutz 1974, Singleton 1977). The
following analysis of the output of 4 of
them was therefore carried out.
Because the American Mosquito Con-
trol Association is still receiving enquiries
about such devices (T. D. Mulhern,
Executive Director, personal communi-
cation) tests of their effectiveness were
made and their potential is discussed in
relation to claims for their mode of ac-
MATERIALS AND METHODS
Three devices were tested both in the
laboratory and in the field. They are
identified as follows: A, "electronic
Vol. 41, No. 4
gadget electronic mosquito repeller,"
supplied by Murray Distributors Ltd.,
North Vancouver B.C.; B, "Electronic
Mosquito Repeller" obtained from Peak
Distributions Ltd., Surrey B. C. and C,
"Moziquit" supplied by Electronic Pest
Controls Ltd., Montreal P.Q. All are
roughly rectangular 6x5x2 cm deep
plastic boxes with a transistor-radio ear-
phone at the bottom right of the front
face. The device claimed to sound like a
male, D, is currently being marketed in
Canada (June 1981) by Canaco Market-
ing, Toronto, Ont. with the name "Anti-
bite." It was apparently also available in
Europe under the name "Antipic." It is
cylindrical, about 7x2.5 cm in diameter
with the earphone facing upward at the
top of the cylinder. The sound level from
each was measured using a Bruel and
Kjaer type 2204 impulse precision sound
level meter with a 4145 condenser mi-
crophone, at a distance of 1 cm. The
amplified output of the meter was fed
into a Tektronix type 3L5 spectrum ana-
lyzer to determine the frequency content
of the sound in the range between 100 Hz
and 50 kHz. An oscillogram of the sound
and its spectrum were photographed si-
multaneously from a Tektronix type 556
dual beam oscilloscope.
The effectiveness of the 4 devices was
investigated using screen cages,
30x30x60 cm high, with a sleeve of 15
cm dia in the Plexiglas front. Eight 1-min
tests were carried out each with 100 pre-
viously untested female Aedes aegypti
(Linnaeus) and five 5-min tests each with
50 untested female Culex pipiens Linnaeus.
The tests on Ae. aegypti were run with
one of the devices on the floor of a cage.
Its sound was transmitted upwards when
it was switched on and it was off in alter-
nate tests. All the mosquitoes landing on
and probing a human hand were counted
during each test. The tests on Cx. pipiens
used a restrained white rat and counts
were made during the 5-min tests of each
device. The counts of Cx. pipiens were
compared with a single control test with
no device in the cage.
Field tests of the devices available in
1977 (A, B and C) were carried out at 6
different sites in mountain woodland
using 5-min counts of landings on 2
seated subjects, each with an exposed
forearm. Alternate counts were made
with the devices switched on and off. Two
counts (first on then off) were made with
device A and 2 with device B. Because
these tests had little effect on biting rates
a final two counts were made with devices
A, B and C switched on together and then
off. The results were analyzed with the x 2
test and the significance determined at
the 1% level (P<0.01). Because of the low
numbers landing, Yates' correction was
applied to the results obtained from Cx.
The output of the 4 commercial devices
is shown in Fig. 1, A to D. The left-hand
traces are accurate representations of 2
msec of the sound from each device in the
audible frequency range. The right-hand
traces are spectral analyses showing a
linear representation of sound pressure
(vertical scale) at frequencies between
and 50 kHz (horizontal scale). The sen-
sitivity of the microphone is gready de-
creased above 20 kHz so that the vertical
calibration does not hold good for the
right half of the spectra. The spectra
show a series of peaks. The peak at zero
frequency is an artifact generated by the
analyzer. To its right is the low-amplitude
fundamental frequency of the device (ar-
rowed) and succeeding peaks are those of
the second (asterisked) and higher har-
monics, each of which is linearly related
to the sound pressure at that frequency.
Each sound had many harmonics at mul-
tiples of its fundamental frequency, re-
lated to the pulsed output of its oscillator.
The less obvious pulsing of device D gives
rise to fewer harmonics and its output in
the audible range is mostly the second
harmonic at 10 kHz. As the distribution
of harmonic peaks indicates, both pitch
and quality (timbre) of all 4 sounds are
audibly different from each other. The
right-hand side of the spectra in Fig. 1
5 10 15 20
Fig. 1. Left traces, oscillograms of 2 msec of the sound from devices A, B, C arid D, re-
spectively. Right traces, spectral analyses of sounds at left; fundamental frequency arrowed,
second harmonic with an asterisk and higher harmonics to the right. The vertical calibration
represents 50 mV RMS output from the spectrum analyzer showing reduced amplification for
traces B and D.
show slight deflections which indicate that
ultrasonic frequencies are present in the
output of all 4 devices.
The sounds are not particularly in-
tense, their pressure, at a distance of 1
cm, and their fundamental and highest
amplitude harmonic frequencies are
shown in Table 1.
Table 1. Characteristics of the sound of the
4 devices tested.
74 dBA 2.0 kHz
Sounds of this intensity are not audible
at distances of more than about 1 m, over
the background noise of a "suburban liv-
ing room" (Broch 1969). Nevertheless
they are about 50- 100 times louder than -
the sound of a female mosquito which is
about 40 dBA at 1 cm (Wishart and Rior-
dan 1959) and they are several hundred
times less intense than that of an
echolocating bat which is about 100 dB at
30 cm (Belton and Kempster 1962).
Between 70 and 87 of the caged Ae.
aegypti landed and were probing the ex-
posed hand at the end of each 1-min test
and the numbers were not significantly
different whether any of the devices was
on or off. Between 3 and 14 of the 50 Cx.
pipiens landed and probed the rat. In the
Vol. 41, No. 4
test of device B, fewer mosquitoes landed
when the sound was on, however, the x 2
value (with Yates' correction) of 2.13 indi-
cates that this could have occurred by
chance in 1 trial out of 10. In the field
tests there were no significant differences
in numbers landing in 5-min periods
whether one or more of the devices was
switched on or off. The landing rates on
one arm were between 2 and 7/min. Mos-
quitoes were mostly Ae, hexodontus Dyar,
with a few Ae. cataphylla Dyar and Culiseta
The acoustic waveforms of devices A, B
and C are pulsed like that of the Bug-
Shoo and are consistent with a unijunc-
tion oscillator. Kutz (1974) and Singleton
(1977), both investigated device B and de-
scribed its output as a sine wave, however,
they may not have used sufficiently pre-
cise equipment to describe the waveform
accurately. The acoustic characteristics of
transistor-radio earphones are variable
even from the same manufacturer and it
is probable that the amplitude and
quality of the sounds of some of the de-
vices vary from unit to unit but this was
not investigated. None of these 3 devices
correspond in frequency with the flight
sounds of female North American mos-
quitoes. Certainly they could not be ex-
pected to work on the principle of repel-
ling females with the same sound that at-
tracts males. Device D, which was de-
signed to mimic the sound of a male mos-
quito, on the hypothesis that it would
repel mated females, produces a fre-
quency 10 or more times higher than that
of any North American male mosquito
tested. A preliminary investigation of the
antennae of females indicates that they
are less sensitive to sound than those of
males and are tuned to a frequency
around 100 Hz, a fiftieth of the funda-
mental frequency of device D. Sound at
the wing beat frequency of the male, loud
enough to be heard by the females, would
probably be as annoying to humans as the
The results of the laboratory and field
tests confirm those of previous inves-
tigators (Garcia etal. 1976, Gorham 1974,
Helson and Wright 1977, Kutz 1974,
Rasnitsyn et al. 1974, Schreck et al. 1977,
Snow 1977). The devices had no consis-
tent repellent effect. The 3 species en-
countered in the field were different
from those investigated in previously
published field tests but were similar in
their lack of response. The unpleasant
nature of the 4 sounds, as perceived by
the human ear, may be due to the pres-
ence of some ultrasound in the output of
all the devices tested but the negative re-
sults of all laboratory and field tests indi-
cate that ultrasound at these levels is, tike
the audible component, not significantly
repellent to mosquitoes. Greenlee's Bug-
Shoo can be tuned to an inaudibly high
frequency and he states that "all mos-
quitoes seem to be repelled by a high fre-
quency, above 10,000 Hz." Mosquitoes
may have mechanoreceptors sensitive to
high frequencies but no physiological re-
search has been done to locate them.
There is, at present, no scientific proof
that the behavior of female mosquitoes is
affected by any sound frequency but this
should not discourage further research
into the effect of ultrasonic frequencies
and also of frequencies lower than that of
the wingbeat, to which the antennae of
female mosquitoes are sensitive.
I can find no scientific basis for any of
the claims made for the mode of action of
any of the repellers that have been
marketed up to the present. On the con-
trary, most observers exposed to high
biting populations of mosquitoes know
that, although they may jostle for posi-
tion, flying or walking mosquitoes do not
repel each other. My experience of
working near swarms of male mosquitoes
is that their sound does not deter females
from biting and in fact there are several
species in which both males and females
are attracted to hosts with no mutual re-
All 4 putative insect repellers were inef-
fective in laboratory and field tests against
the aforementioned mosquito species al-
though their frequencies extended well
into the ultrasonic range. Because of their
high fundamental frequency, hone of
them could work, as Greenlee (1970) sug-
gested they do, by reproducing a sound
that attracts male mosquitoes. Only one
sample of each device was tested.
Acknowledgment. I thank Dr. H. R.
MacCarthy for his valuable comments on
Belton, P. 1967. Trapping mosquitoes with
sound. Proc. Calif. Mosq. Contr. Assoc.
Belton, P. and R. A. Costello. 1979. Flight
sounds of the females of some mosquitoes of
western Canada. Entomol. Exp. Appl.
Belton, P. and R. H, Kempster. 1962. A field
test on the use of sound to repel the Euro-
pean corn borer. Entomol. Ekd. Appl.
Broch, J, T. 1969. Acoustic noise mea-
surements. Bruel and Kjaer, Copenhagen.
Costello, R. A. and R. A. Brust. 1976. Elec-
trocuters and electronic mosquito repellers.
CCBF Newsletter, Agric. Canada, 2:11-13.
Garcia, R., B. Des Rochers and W. G. Voight,
1 976. Evaluation of electronic mosquito re-
pellers under laboratory and field condi-
tions. Vector Views 23:21-23.
Gorham, J. R. 1974. Tests of mosquito repel-
lents in Alaska. Mosq. News 34:409-415.
Greenlee, L. E. 1970. Build the bug-shoo.
Popular Electronics. July pp. 27-30.
Helson, B. V. and R. E. Wright. 1977, Field
evaluation of electronic mosquito-repellers
in Ontario. Proc. Entomol. Soc. Ont.
Kutz, F. W. 1974. Evaluations of an electronic
mosquito repelling device. Mosq. News
Rasnitsyn, S. P., A. N, Alekseev, R. M. Gor-
nostaeva, E. S. Kupriyanova, A. A. Potapov
and O. V. Razumova. 1974. Negative results
of tests of sound generator specimens de-
signed for mosquito repellence. Med.
Parasitoi. Parasit. Bolezn. 43:706-708. (In
Russian, English summary).
Roth, L. M. 1948. A study of mosquito behav-
ior. An experimental study of the sexual be-
havior of Aedesaegypti (Linnaeus). Am. Midi.
Schreck, C. E., D. E. Weidhaas and C. N.
Smith. 1977. Evaluation of electronic sound
producing devices against Aedes taeniorhyn-
chus and Ae. soilkitans. Mosq. News 37-529-
Singleton, R. E. 1977. Evaluation of two
mosquito-repelling devices. Mosq, NeWs
Snow, W. F. 1977. Trials with a mosquito-
repelling device in West Africa. Trans. R.
Soc. Trop. Med. Hyg. 71:449-450.
Wishart, G. and D. F. Riordan. 1959. Flight
responses to various sounds by adult males
of Aeies aegypti (L.) (Diptera: Culicidae).
Canad. Entomol. 91:181-191.